Step 1 · AccessFictional example: Felsbach Armaturen
Whoever locks out crawlers locks out the answer.
AI search engines fetch pages with their own crawlers before they can quote from them. Many websites block these crawlers wholesale in robots.txt, mostly out of concern about training use. The block then also hits retrieval for answers: what is not fetched cannot be quoted.
A robots.txt is a request, not a barrier. Whether a crawler honours it is up to its operator. And which crawler feeds which answer surface changes; that is why the file in the example carries a date, and the check repeats with every measurement.
Block deliberately, not wholesale. Anyone who excludes a crawler should know which answers they lose by doing so. In the example, Felsbach blocks one of three crawlers without noticing.
felsbach.example/robots.txt · as of September 2026fictional
User-agent: * Allow: / User-agent: PerplexityBot Disallow: / User-agent: GPTBot Allow: / User-agent: ClaudeBot Allow: /
Access · crawlers in the example
- GPTBotallowed
- ClaudeBotallowed
- PerplexityBotblocked, Disallow: /
2 of 3 crawlers allowed. Example names; which crawler feeds which answer surface is not stated here.